TCS Daily


'Just-in-Case': How to Think About Uncertainty and Global Warming

By Arnold Kling - February 14, 2007 12:00 AM

"Let's just say that global warming deniers are now on a par with Holocaust deniers, though one denies the past and the other denies the present and future."
-- Ellen Goodman

"Global warming is a false myth and every serious person and scientist says so. It is not fair to refer to the U.N. panel. IPCC is not a scientific institution: it's a political body, a sort of non-government organization of green flavor. It's neither a forum of neutral scientists nor a balanced group of scientists. These people are politicized scientists who arrive there with a one-sided opinion and a one-sided assignment. Also, it's an undignified slapstick that people don't wait for the full report in May 2007 but instead respond, in such a serious way, to the summary for policymakers where all the "but's" are scratched, removed, and replaced by oversimplified theses. This is clearly such an incredible failure of so many people, from journalists to politicians."
-- Czech President Vaclav Klaus

Pundits, politicians, and the public have a hard time coming to grips with uncertainty. This makes the atmosphere for debating global warming policy especially foul, because the key issues with global warming are the uncertainties involved. Those who would try to reduce the issue of global warming to a yes-or-no question ("do you believe or do you deny?") are not scientists.

Real scientists understand uncertainty. Real science deals with uncertainty through relentless, skeptical inquiry. Real science resolves arguments not with consensus, but with data.

My understanding of global warming is influenced by my background in applied statistics and economics. There certainly are scientists who have spent more time than I have analyzing the meteorological data. However, before you call me a "hack," make sure that you are capable of understanding, say, Martin Weitzman's critique of the Stern report.

The Known and the Unknown

Two facts are known. One is that the level of atmospheric carbon dioxide has been climbing exponentially. In fact, the overall level of human activity, as measured by total Gross Domestic Product, is perhaps 50 times higher than it was one hundred years ago.

Another fact is that over the past 30 years, the average global temperature has increased in total by between 0.5 degrees and one degree centigrade. That means that the average annual rate of increase has been less than one-tenth of one degree per year.

The global warming that has taken place so far is minor. The improvement in living standards that has taken place in the past one hundred years is enormous.

The global warming issue has nothing to do with the global warming that has taken place to date. It has everything to do with the global warming that will take place in the future. This is a matter that depends on climate forecasting.

The climate forecasting challenge is to predict the rate of change of global temperature, based on the level of atmospheric carbon dioxide. From a purely statistical standpoint, we do not have the data to do this. The current level of atmospheric carbon dioxide is quite high, even relative to a quarter century ago. We probably have not had enough experience to know the rate of change in global temperature that will result from this high level. The attempts to cope with this ignorance are known as climate models. Given the sheer complexity of global climate, it is impossible to have much confidence in these models. Perhaps they are better than nothing -- that is, better than a simple statistical extrapolation. However, my guess would be otherwise.

For a decade, beginning in about 1978, the economics profession witnessed the utter discrediting of a similar exercise in complex modeling -- so-called macroeconometrics. Macroeconometrics was supposed to enable you to do something like predict nominal GDP (GDP not adjusted for inflation) based on the money supply.

Looking at historical data, this almost seems too easy -- the level of GDP and the level of the money supply both rise over time. Economists thought that they might remove this spurious correlation by "de-trending" the data and then looking at the correlation adjusted for this trend. Unfortunately -- and this is what began to hit home in the late 1970's -- "de-trending" did not remove the spurious correlation. There was a lot of highly technical analysis undertaken, including work by Nobel Laureate Clive Granger. My take-away from this work is that the amount of meaningful information in macroeconomic data is much less than is necessary for macroeconometric models to be successful. Simple statistical extrapolations work just about as well, and often better.

Climate data looks to me suspiciously like macroeconomic data. The true information content probably is not sufficient to produce a reliable model for forecasting.

"Just in Case"

If I thought that climate models were highly accurate, then I would be opposed to any major near-term policy to address global warming. The reason is that the climate models predict gradual, modest global warming over the next century. What that means is that, relative to future GDP, the costs are tolerable. As Weitzman points out, the mainstream economic approach to trading off tolerable future costs against current costs would argue against making significant sacrifices today.

However, Weitzman implicitly shares my concern with climate models. Obviously, we have nothing to worry about if the models are too pessimistic. If it turns out that over the next decade global temperatures edge down, or rise more slowly than the models predict, then we will be relieved.

The troublesome possibility is that the models are not pessimistic enough. In fact, Weitzman would argue, and I concur, that the case for doing something today about global warming rests on the fear of the scenario of accelerated near-term climate change -- increases in temperature at a rate that is on the high end of the range being forecast by climate models.

The ideal approach would be a "just-in-case" climate-change mitigation plan. If global warming stays at or under current baseline projections, we probably would do best to simply just adapt. However, if global warming accelerates, we would want to take strong steps to counteract it.

In that sense, climate engineering, in which we attempt to use technological means to manipulate global temperatures, would be a good fit. If we had such mechanisms, we could use them on an as-needed basis.

Martin Weitzman also discusses the issue of "just-in-case" policies.

"...many hard questions need to be asked. What are early-warning signs of impending environmental disasters like melting ice sheets or thermohaline inversions? How much would it cost to put sensors in place that might detect early-warning signals of impending climatic catastrophes? How early might the warning be before the full effects are felt? What could we do as an emergency response if we received such an early-warning signal? Would last-ditch emergency measures help to ward off disaster by reversing the worst consequences of global warming in time?"

Weitzman concedes that the "just-in-case" approach may not be feasible.

"It may well turn out that the option value of waiting for better information about catastrophic tail events is negligible (because early detection is impossible, or it is too expensive, or it comes too late, or because nothing practical can be done about undoing greenhouse warming anyway), but these are conclusions we need to reach empirically, rather than postulating them initially."

Stripped of jargon, what Weitzman is saying is that we should try to study whether a just-in-case approach can work, rather than limit ourselves to the choices of either making huge economic sacrifices or running the risk of climate catastrophe.

If we lack "just-in-case" mechanisms, then any approach that we take toward climate change risks making significant errors. We might sacrifice a lot of the world's standard of living in order to reduce carbon dioxide emissions, only to discover that it was unnecessary, because global warming was not going to accelerate, regardless. Conversely, the reductions that we carry out might turn out to be insufficient.

If we do opt for sacrifice, then most economists would agree with Weitzman that the best approach would be to tax carbon emissions.

"One can only wish that U.S. political leaders might have the insight to understand and the courage to act upon the breathtakingly-simple market-friendly idea that the right carbon tax could do way more to unleash the decentralized power of greedy, self-seeking American inventive genius on the problem of developing economically-feasible non-carbon-intensive alternative technologies than all of the command-and-control schemes and patchwork subsidies making the rounds in Washington these days."

It is possible to have a civilized, sensible discussion about the issue of global climate change. However, doing so requires speaking in the language of uncertainty, rather than moral righteousness.

Arnold Kling is author of Learning Economics and Crisis of Abundance.

Categories:
|

TCS Daily Archives